Importing Libraries

import re
import requests
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from tensorflow.keras import Sequential
from tensorflow.keras.layers import Dense,Dropout
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix,classification_report
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix,classification_report
import matplotlib.pyplot as plt
from sklearn import preprocessing 

1. Loading Dataset

def load_crime_data(fileName):
    
    r = requests.get('http://archive.ics.uci.edu/ml/machine-learning-databases/communities/communities.names')
    all_columns = re.findall(r'\@attribute (\w*)', r.text)
    del r

    output_columns = ['ViolentCrimesPerPop']
    
    data = pd.read_csv(fileName, names=all_columns, index_col=False)
    
    return output_columns,all_columns,data

output_columns,all_columns,data  = load_crime_data('data\communities.data')

2. EDA

data.head(10)

	state	county	community	communityname	fold	population	householdsize	racepctblack	racePctWhite	racePctAsian	...	LandArea	PopDens	PctUsePubTrans	PolicCars	PolicOperBudg	LemasPctPolicOnPatr	LemasGangUnitDeploy	LemasPctOfficDrugUn	PolicBudgPerPop	ViolentCrimesPerPop
0	8	?	?	Lakewoodcity	1	0.19	0.33	0.02	0.90	0.12	...	0.12	0.26	0.20	0.06	0.04	0.9	0.5	0.32	0.14	0.20
1	53	?	?	Tukwilacity	1	0.00	0.16	0.12	0.74	0.45	...	0.02	0.12	0.45	?	?	?	?	0.00	?	0.67
2	24	?	?	Aberdeentown	1	0.00	0.42	0.49	0.56	0.17	...	0.01	0.21	0.02	?	?	?	?	0.00	?	0.43
3	34	5	81440	Willingborotownship	1	0.04	0.77	1.00	0.08	0.12	...	0.02	0.39	0.28	?	?	?	?	0.00	?	0.12
4	42	95	6096	Bethlehemtownship	1	0.01	0.55	0.02	0.95	0.09	...	0.04	0.09	0.02	?	?	?	?	0.00	?	0.03
5	6	?	?	SouthPasadenacity	1	0.02	0.28	0.06	0.54	1.00	...	0.01	0.58	0.10	?	?	?	?	0.00	?	0.14
6	44	7	41500	Lincolntown	1	0.01	0.39	0.00	0.98	0.06	...	0.05	0.08	0.06	?	?	?	?	0.00	?	0.03
7	6	?	?	Selmacity	1	0.01	0.74	0.03	0.46	0.20	...	0.01	0.33	0.00	?	?	?	?	0.00	?	0.55
8	21	?	?	Hendersoncity	1	0.03	0.34	0.20	0.84	0.02	...	0.04	0.17	0.04	?	?	?	?	0.00	?	0.53
9	29	?	?	Claytoncity	1	0.01	0.40	0.06	0.87	0.30	...	0.00	0.47	0.11	?	?	?	?	0.00	?	0.15

10 rows × 128 columns

data.describe()

	state	fold	population	householdsize	racepctblack	racePctWhite	racePctAsian	racePctHisp	agePct12t21	agePct12t29	...	PctForeignBorn	PctBornSameState	PctSameHouse85	PctSameCity85	PctSameState85	LandArea	PopDens	PctUsePubTrans	LemasPctOfficDrugUn	ViolentCrimesPerPop
count	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	...	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000	1994.000000
mean	28.683551	5.493982	0.057593	0.463395	0.179629	0.753716	0.153681	0.144022	0.424218	0.493867	...	0.215552	0.608892	0.535050	0.626424	0.651530	0.065231	0.232854	0.161685	0.094052	0.237979
std	16.397553	2.873694	0.126906	0.163717	0.253442	0.244039	0.208877	0.232492	0.155196	0.143564	...	0.231134	0.204329	0.181352	0.200521	0.198221	0.109459	0.203092	0.229055	0.240328	0.232985
min	1.000000	1.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	...	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
25%	12.000000	3.000000	0.010000	0.350000	0.020000	0.630000	0.040000	0.010000	0.340000	0.410000	...	0.060000	0.470000	0.420000	0.520000	0.560000	0.020000	0.100000	0.020000	0.000000	0.070000
50%	34.000000	5.000000	0.020000	0.440000	0.060000	0.850000	0.070000	0.040000	0.400000	0.480000	...	0.130000	0.630000	0.540000	0.670000	0.700000	0.040000	0.170000	0.070000	0.000000	0.150000
75%	42.000000	8.000000	0.050000	0.540000	0.230000	0.940000	0.170000	0.160000	0.470000	0.540000	...	0.280000	0.777500	0.660000	0.770000	0.790000	0.070000	0.280000	0.190000	0.000000	0.330000
max	56.000000	10.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	...	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000

8 rows × 102 columns

# data.hist()

data.isna().sum()

state                   0
county                  0
community               0
communityname           0
fold                    0
population              0
householdsize           0
racepctblack            0
racePctWhite            0
racePctAsian            0
racePctHisp             0
agePct12t21             0
agePct12t29             0
agePct16t24             0
agePct65up              0
numbUrban               0
pctUrban                0
medIncome               0
pctWWage                0
pctWFarmSelf            0
pctWInvInc              0
pctWSocSec              0
pctWPubAsst             0
pctWRetire              0
medFamInc               0
perCapInc               0
whitePerCap             0
blackPerCap             0
indianPerCap            0
AsianPerCap             0
                       ..
PctSameHouse85          0
PctSameCity85           0
PctSameState85          0
LemasSwornFT            0
LemasSwFTPerPop         0
LemasSwFTFieldOps       0
LemasSwFTFieldPerPop    0
LemasTotalReq           0
LemasTotReqPerPop       0
PolicReqPerOffic        0
PolicPerPop             0
RacialMatchCommPol      0
PctPolicWhite           0
PctPolicBlack           0
PctPolicHisp            0
PctPolicAsian           0
PctPolicMinor           0
OfficAssgnDrugUnits     0
NumKindsDrugsSeiz       0
PolicAveOTWorked        0
LandArea                0
PopDens                 0
PctUsePubTrans          0
PolicCars               0
PolicOperBudg           0
LemasPctPolicOnPatr     0
LemasGangUnitDeploy     0
LemasPctOfficDrugUn     0
PolicBudgPerPop         0
ViolentCrimesPerPop     0
Length: 128, dtype: int64

data = data.replace('?', np.nan)

data.isna().sum()

state                      0
county                  1174
community               1177
communityname              0
fold                       0
population                 0
householdsize              0
racepctblack               0
racePctWhite               0
racePctAsian               0
racePctHisp                0
agePct12t21                0
agePct12t29                0
agePct16t24                0
agePct65up                 0
numbUrban                  0
pctUrban                   0
medIncome                  0
pctWWage                   0
pctWFarmSelf               0
pctWInvInc                 0
pctWSocSec                 0
pctWPubAsst                0
pctWRetire                 0
medFamInc                  0
perCapInc                  0
whitePerCap                0
blackPerCap                0
indianPerCap               0
AsianPerCap                0
                        ... 
PctSameHouse85             0
PctSameCity85              0
PctSameState85             0
LemasSwornFT            1675
LemasSwFTPerPop         1675
LemasSwFTFieldOps       1675
LemasSwFTFieldPerPop    1675
LemasTotalReq           1675
LemasTotReqPerPop       1675
PolicReqPerOffic        1675
PolicPerPop             1675
RacialMatchCommPol      1675
PctPolicWhite           1675
PctPolicBlack           1675
PctPolicHisp            1675
PctPolicAsian           1675
PctPolicMinor           1675
OfficAssgnDrugUnits     1675
NumKindsDrugsSeiz       1675
PolicAveOTWorked        1675
LandArea                   0
PopDens                    0
PctUsePubTrans             0
PolicCars               1675
PolicOperBudg           1675
LemasPctPolicOnPatr     1675
LemasGangUnitDeploy     1675
LemasPctOfficDrugUn        0
PolicBudgPerPop         1675
ViolentCrimesPerPop        0
Length: 128, dtype: int64

len(data['ViolentCrimesPerPop'])

1994

for i in range(len(data['ViolentCrimesPerPop'])):
    if data['ViolentCrimesPerPop'][i] > 0.1:
        data['ViolentCrimesPerPop'][i] = 1  # High Crime
    else:
        data['ViolentCrimesPerPop'][i] = 0 # Low Crime Rate

c:\users\sahil\appdata\local\programs\python\python36\lib\site-packages\ipykernel_launcher.py:3: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  This is separate from the ipykernel package so we can avoid doing imports until
c:\users\sahil\appdata\local\programs\python\python36\lib\site-packages\ipykernel_launcher.py:5: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  """

# Calculating the percentage of positive and negative instances in the dataset
percentage_intances = data.groupby('ViolentCrimesPerPop').size() * 100 / len(data)
print(percentage_intances)
print("------------------")
print("Percentage Positive Instance = {}\nPercentage Negative Instance = {} ".format(percentage_intances[1],percentage_intances[0]))

ViolentCrimesPerPop
0.0    37.261785
1.0    62.738215
dtype: float64
------------------
Percentage Positive Instance = 62.7382146439318
Percentage Negative Instance = 37.2617853560682 

data.columns

Index(['state', 'county', 'community', 'communityname', 'fold', 'population',
       'householdsize', 'racepctblack', 'racePctWhite', 'racePctAsian',
       ...
       'LandArea', 'PopDens', 'PctUsePubTrans', 'PolicCars', 'PolicOperBudg',
       'LemasPctPolicOnPatr', 'LemasGangUnitDeploy', 'LemasPctOfficDrugUn',
       'PolicBudgPerPop', 'ViolentCrimesPerPop'],
      dtype='object', length=128)

# Non-predictive Categorical Variables 
for col in ['state','county','community','communityname']:
    data[col].fillna(0,inplace=True)

(data.isnull().sum()).tail()

LemasPctPolicOnPatr    1675
LemasGangUnitDeploy    1675
LemasPctOfficDrugUn       0
PolicBudgPerPop        1675
ViolentCrimesPerPop       0
dtype: int64

# Filing null value with mean value

data = data.apply(lambda x:x.fillna(x.value_counts().index[0]))

data.head()

	state	county	community	communityname	fold	population	householdsize	racepctblack	racePctWhite	racePctAsian	...	LandArea	PopDens	PctUsePubTrans	PolicCars	PolicOperBudg	LemasPctPolicOnPatr	LemasGangUnitDeploy	LemasPctOfficDrugUn	PolicBudgPerPop	ViolentCrimesPerPop
0	8	0	0	Lakewoodcity	1	0.19	0.33	0.02	0.90	0.12	...	0.12	0.26	0.20	0.06	0.04	0.9	0.5	0.32	0.14	1.0
1	53	0	0	Tukwilacity	1	0.00	0.16	0.12	0.74	0.45	...	0.02	0.12	0.45	0.02	0.02	0.84	0	0.00	0.12	1.0
2	24	0	0	Aberdeentown	1	0.00	0.42	0.49	0.56	0.17	...	0.01	0.21	0.02	0.02	0.02	0.84	0	0.00	0.12	1.0
3	34	5	81440	Willingborotownship	1	0.04	0.77	1.00	0.08	0.12	...	0.02	0.39	0.28	0.02	0.02	0.84	0	0.00	0.12	1.0
4	42	95	6096	Bethlehemtownship	1	0.01	0.55	0.02	0.95	0.09	...	0.04	0.09	0.02	0.02	0.02	0.84	0	0.00	0.12	0.0

5 rows × 128 columns

print("Missing values after Imputing : ", data.isnull().sum().sum())

Missing values after Imputing :  0

data = data.drop(['state','county','community','communityname','fold'],axis=1)

len(data.columns)

123

data.to_csv('data/clean_data.csv', encoding='utf-8')

corr_matrix = data.corr()

fig, ax = plt.subplots(figsize=(50,50))
ax = sns.heatmap(corr_matrix,
                annot=True,
                linewidths=0.5,
                fmt=".2f", 
                cmap = "YlGnBu");

all_col = data.columns

input_columns = all_col[:-1]

len(input_columns)

122

3. Splitting data (Training / Testing / Validation )

X = data[input_columns]
Y = data['ViolentCrimesPerPop']

X.head()

	population	householdsize	racepctblack	racePctWhite	racePctAsian	racePctHisp	agePct12t21	agePct12t29	agePct16t24	agePct65up	...	PolicAveOTWorked	LandArea	PopDens	PctUsePubTrans	PolicCars	PolicOperBudg	LemasPctPolicOnPatr	LemasGangUnitDeploy	LemasPctOfficDrugUn	PolicBudgPerPop
0	0.19	0.33	0.02	0.90	0.12	0.17	0.34	0.47	0.29	0.32	...	0.29	0.12	0.26	0.20	0.06	0.04	0.9	0.5	0.32	0.14
1	0.00	0.16	0.12	0.74	0.45	0.07	0.26	0.59	0.35	0.27	...	0.19	0.02	0.12	0.45	0.02	0.02	0.84	0	0.00	0.12
2	0.00	0.42	0.49	0.56	0.17	0.04	0.39	0.47	0.28	0.32	...	0.19	0.01	0.21	0.02	0.02	0.02	0.84	0	0.00	0.12
3	0.04	0.77	1.00	0.08	0.12	0.10	0.51	0.50	0.34	0.21	...	0.19	0.02	0.39	0.28	0.02	0.02	0.84	0	0.00	0.12
4	0.01	0.55	0.02	0.95	0.09	0.05	0.38	0.38	0.23	0.36	...	0.19	0.04	0.09	0.02	0.02	0.02	0.84	0	0.00	0.12

5 rows × 122 columns

Y.head()

0    1.0
1    1.0
2    1.0
3    1.0
4    0.0
Name: ViolentCrimesPerPop, dtype: float64

Normalization

Y = Y.values.reshape((Y.shape[0],1))
X = preprocessing.normalize(X)
Y = preprocessing.normalize(Y)
Y = Y.reshape((Y.shape[0],))

from collections import Counter
print(Counter(Y).keys())
print(Counter(Y).values())

dict_keys([1.0, 0.0])
dict_values([1251, 743])

X

array([[0.04108821, 0.07136374, 0.00432508, ..., 0.10812688, 0.0692012 ,
        0.03027553],
       [0.        , 0.03520005, 0.02640004, ..., 0.        , 0.        ,
        0.02640004],
       [0.        , 0.09842032, 0.11482371, ..., 0.        , 0.        ,
        0.02812009],
       ...,
       [0.03465583, 0.08014161, 0.05414973, ..., 0.        , 0.19710503,
        0.0606477 ],
       [0.0161467 , 0.10293519, 0.01211002, ..., 0.        , 0.04440342,
        0.03633007],
       [0.03685304, 0.14372686, 0.02579713, ..., 0.0921326 , 0.18426521,
        0.02395448]])

def train_val_test_split(X, y, train_size, val_size, test_size):
    X_train_val, X_test, y_train_val, y_test = train_test_split(X, y, test_size = test_size)
    relative_train_size = train_size / (val_size + train_size)
    X_train, X_val, y_train, y_val = train_test_split(X_train_val, y_train_val,
                                                      train_size = relative_train_size, test_size = 1-relative_train_size)
    return X_train, X_val, X_test, y_train, y_val, y_test

X_train, X_val, X_test, y_train, y_val, y_test = train_val_test_split(X, Y, 0.7, 0.1, 0.2)

len(X_train), len(X_val), len(X_test), len(y_train), len(y_val), len(y_test)

(1395, 200, 399, 1395, 200, 399)

4. StandardScaler

# from sklearn.preprocessing import StandardScaler

# ss=StandardScaler()
# X_train=ss.fit_transform(X_train)
# X_test=ss.fit_transform(X_test)
# X_val=ss.fit_transform(X_val)

X_train

array([[0.00201209, 0.11670138, 0.00201209, ..., 0.        , 0.        ,
        0.02414511],
       [0.00663772, 0.05973948, 0.03540118, ..., 0.        , 0.        ,
        0.02655088],
       [0.00411675, 0.10497717, 0.00617513, ..., 0.        , 0.        ,
        0.02470051],
       ...,
       [0.00624917, 0.1374818 , 0.00624917, ..., 0.        , 0.        ,
        0.02499669],
       [0.00208557, 0.12930508, 0.00208557, ..., 0.        , 0.        ,
        0.02502679],
       [0.00471498, 0.15559423, 0.02043157, ..., 0.        , 0.        ,
        0.01885991]])

5. Creating Neural Network Model

#Creating a pipeline
model = Sequential()

#1st hidden layer with input layer
model.add(Dense(units=500,activation="relu",input_dim=122))

#2nd hidden layer
model.add(Dense(units=300,activation="relu",))

#3rd hidden layer
model.add(Dense(units=100,activation="relu",))

#output layer
model.add(Dense(units=1,activation="sigmoid"))

model.summary()

Model: "sequential_8"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_32 (Dense)             (None, 500)               61500     
_________________________________________________________________
dense_33 (Dense)             (None, 300)               150300    
_________________________________________________________________
dense_34 (Dense)             (None, 100)               30100     
_________________________________________________________________
dense_35 (Dense)             (None, 1)                 101       
=================================================================
Total params: 242,001
Trainable params: 242,001
Non-trainable params: 0
_________________________________________________________________

6. Compiling Model

model.compile(optimizer="adam",loss="binary_crossentropy",metrics=["accuracy"])

7. Training Model

model_his=model.fit(X_train,y_train, validation_data=(X_val, y_val), batch_size=16,epochs=100,verbose=1)

Epoch 1/100
88/88 [==============================] - 1s 7ms/step - loss: 0.4896 - accuracy: 0.7391 - val_loss: 0.3887 - val_accuracy: 0.8100
Epoch 2/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3912 - accuracy: 0.8158 - val_loss: 0.4112 - val_accuracy: 0.7800
Epoch 3/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3870 - accuracy: 0.8194 - val_loss: 0.3789 - val_accuracy: 0.8100
Epoch 4/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3651 - accuracy: 0.8272 - val_loss: 0.3925 - val_accuracy: 0.8050
Epoch 5/100
88/88 [==============================] - 0s 6ms/step - loss: 0.3689 - accuracy: 0.8308 - val_loss: 0.3664 - val_accuracy: 0.8350
Epoch 6/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3562 - accuracy: 0.8315 - val_loss: 0.3702 - val_accuracy: 0.8350
Epoch 7/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3480 - accuracy: 0.8387 - val_loss: 0.3570 - val_accuracy: 0.8400
Epoch 8/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3541 - accuracy: 0.8459 - val_loss: 0.3508 - val_accuracy: 0.8400
Epoch 9/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3625 - accuracy: 0.8366 - val_loss: 0.3569 - val_accuracy: 0.8350
Epoch 10/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3379 - accuracy: 0.8452 - val_loss: 0.3573 - val_accuracy: 0.8250
Epoch 11/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3480 - accuracy: 0.8308 - val_loss: 0.4157 - val_accuracy: 0.8000
Epoch 12/100
88/88 [==============================] - 0s 6ms/step - loss: 0.3453 - accuracy: 0.8394 - val_loss: 0.4400 - val_accuracy: 0.8050
Epoch 13/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3438 - accuracy: 0.8387 - val_loss: 0.4112 - val_accuracy: 0.7900
Epoch 14/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3397 - accuracy: 0.8344 - val_loss: 0.3906 - val_accuracy: 0.8200
Epoch 15/100
88/88 [==============================] - 0s 4ms/step - loss: 0.3315 - accuracy: 0.8480 - val_loss: 0.3529 - val_accuracy: 0.8400
Epoch 16/100
88/88 [==============================] - 0s 3ms/step - loss: 0.3339 - accuracy: 0.8437 - val_loss: 0.3543 - val_accuracy: 0.8400
Epoch 17/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3279 - accuracy: 0.8495 - val_loss: 0.3512 - val_accuracy: 0.8400
Epoch 18/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3290 - accuracy: 0.8423 - val_loss: 0.3486 - val_accuracy: 0.8500
Epoch 19/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3271 - accuracy: 0.8437 - val_loss: 0.3543 - val_accuracy: 0.8350
Epoch 20/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3165 - accuracy: 0.8516 - val_loss: 0.3584 - val_accuracy: 0.8400
Epoch 21/100
88/88 [==============================] - 0s 4ms/step - loss: 0.3354 - accuracy: 0.8416 - val_loss: 0.3828 - val_accuracy: 0.8250
Epoch 22/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3150 - accuracy: 0.8444 - val_loss: 0.3587 - val_accuracy: 0.8350
Epoch 23/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3196 - accuracy: 0.8495 - val_loss: 0.4433 - val_accuracy: 0.8150
Epoch 24/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3118 - accuracy: 0.8480 - val_loss: 0.3581 - val_accuracy: 0.8300
Epoch 25/100
88/88 [==============================] - 0s 4ms/step - loss: 0.3323 - accuracy: 0.8480 - val_loss: 0.4168 - val_accuracy: 0.8300
Epoch 26/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3258 - accuracy: 0.8416 - val_loss: 0.4003 - val_accuracy: 0.8350
Epoch 27/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3154 - accuracy: 0.8509 - val_loss: 0.4062 - val_accuracy: 0.8300
Epoch 28/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3024 - accuracy: 0.8545 - val_loss: 0.3960 - val_accuracy: 0.8250
Epoch 29/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3085 - accuracy: 0.8502 - val_loss: 0.3599 - val_accuracy: 0.8350
Epoch 30/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3086 - accuracy: 0.8430 - val_loss: 0.3669 - val_accuracy: 0.8350
Epoch 31/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2929 - accuracy: 0.8588 - val_loss: 0.3806 - val_accuracy: 0.8400
Epoch 32/100
88/88 [==============================] - 0s 5ms/step - loss: 0.3058 - accuracy: 0.8523 - val_loss: 0.4114 - val_accuracy: 0.8300
Epoch 33/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2919 - accuracy: 0.8530 - val_loss: 0.4182 - val_accuracy: 0.8350
Epoch 34/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2968 - accuracy: 0.8538 - val_loss: 0.3743 - val_accuracy: 0.8250
Epoch 35/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2863 - accuracy: 0.8595 - val_loss: 0.4106 - val_accuracy: 0.8350
Epoch 36/100
88/88 [==============================] - 1s 6ms/step - loss: 0.2846 - accuracy: 0.8624 - val_loss: 0.4202 - val_accuracy: 0.8450
Epoch 37/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2892 - accuracy: 0.8566 - val_loss: 0.5367 - val_accuracy: 0.8300
Epoch 38/100
88/88 [==============================] - 1s 6ms/step - loss: 0.3040 - accuracy: 0.8480 - val_loss: 0.4303 - val_accuracy: 0.8300
Epoch 39/100
88/88 [==============================] - 0s 6ms/step - loss: 0.2835 - accuracy: 0.8652 - val_loss: 0.4013 - val_accuracy: 0.8400
Epoch 40/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2794 - accuracy: 0.8638 - val_loss: 0.4582 - val_accuracy: 0.8350
Epoch 41/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2685 - accuracy: 0.8616 - val_loss: 0.4657 - val_accuracy: 0.8250
Epoch 42/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2682 - accuracy: 0.8602 - val_loss: 0.5135 - val_accuracy: 0.8350
Epoch 43/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2833 - accuracy: 0.8545 - val_loss: 0.3944 - val_accuracy: 0.8100
Epoch 44/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2695 - accuracy: 0.8588 - val_loss: 0.4710 - val_accuracy: 0.8050
Epoch 45/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2816 - accuracy: 0.8559 - val_loss: 0.4496 - val_accuracy: 0.8150
Epoch 46/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2557 - accuracy: 0.8738 - val_loss: 0.4982 - val_accuracy: 0.8250
Epoch 47/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2585 - accuracy: 0.8695 - val_loss: 0.6212 - val_accuracy: 0.8200
Epoch 48/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2723 - accuracy: 0.8681 - val_loss: 0.4059 - val_accuracy: 0.8300
Epoch 49/100
88/88 [==============================] - 1s 6ms/step - loss: 0.2489 - accuracy: 0.8746 - val_loss: 0.6066 - val_accuracy: 0.8250
Epoch 50/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2500 - accuracy: 0.8710 - val_loss: 0.4922 - val_accuracy: 0.8300
Epoch 51/100
88/88 [==============================] - 0s 3ms/step - loss: 0.2574 - accuracy: 0.8717 - val_loss: 0.4756 - val_accuracy: 0.8300
Epoch 52/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2317 - accuracy: 0.8824 - val_loss: 0.6394 - val_accuracy: 0.8500
Epoch 53/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2550 - accuracy: 0.8717 - val_loss: 0.4887 - val_accuracy: 0.8250
Epoch 54/100
88/88 [==============================] - 0s 6ms/step - loss: 0.2276 - accuracy: 0.8918 - val_loss: 0.6030 - val_accuracy: 0.8400
Epoch 55/100
88/88 [==============================] - 1s 6ms/step - loss: 0.2402 - accuracy: 0.8796 - val_loss: 0.5038 - val_accuracy: 0.8200
Epoch 56/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2328 - accuracy: 0.8867 - val_loss: 0.5570 - val_accuracy: 0.8200
Epoch 57/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2349 - accuracy: 0.8839 - val_loss: 0.5470 - val_accuracy: 0.8050
Epoch 58/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2275 - accuracy: 0.8910 - val_loss: 0.6889 - val_accuracy: 0.8250
Epoch 59/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2287 - accuracy: 0.8903 - val_loss: 0.5728 - val_accuracy: 0.8250
Epoch 60/100
88/88 [==============================] - 1s 7ms/step - loss: 0.2176 - accuracy: 0.9032 - val_loss: 0.6534 - val_accuracy: 0.8200
Epoch 61/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2084 - accuracy: 0.9039 - val_loss: 0.6086 - val_accuracy: 0.8300
Epoch 62/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2220 - accuracy: 0.8925 - val_loss: 0.5996 - val_accuracy: 0.8150
Epoch 63/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2271 - accuracy: 0.8903 - val_loss: 0.6477 - val_accuracy: 0.8300
Epoch 64/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2099 - accuracy: 0.8961 - val_loss: 0.6047 - val_accuracy: 0.8150
Epoch 65/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2126 - accuracy: 0.9032 - val_loss: 0.7487 - val_accuracy: 0.8100
Epoch 66/100
88/88 [==============================] - 1s 6ms/step - loss: 0.2381 - accuracy: 0.8803 - val_loss: 0.6297 - val_accuracy: 0.8200
Epoch 67/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2034 - accuracy: 0.9061 - val_loss: 0.6093 - val_accuracy: 0.8000
Epoch 68/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1882 - accuracy: 0.9090 - val_loss: 0.7932 - val_accuracy: 0.8050
Epoch 69/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1894 - accuracy: 0.9133 - val_loss: 0.7264 - val_accuracy: 0.8200
Epoch 70/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1795 - accuracy: 0.9168 - val_loss: 0.8564 - val_accuracy: 0.7950
Epoch 71/100
88/88 [==============================] - 0s 5ms/step - loss: 0.2100 - accuracy: 0.9018 - val_loss: 0.7264 - val_accuracy: 0.8300
Epoch 72/100
88/88 [==============================] - 1s 6ms/step - loss: 0.2032 - accuracy: 0.8982 - val_loss: 0.6263 - val_accuracy: 0.7850
Epoch 73/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2475 - accuracy: 0.8853 - val_loss: 0.5125 - val_accuracy: 0.8200
Epoch 74/100
88/88 [==============================] - 0s 4ms/step - loss: 0.2080 - accuracy: 0.9039 - val_loss: 0.6539 - val_accuracy: 0.8150
Epoch 75/100
88/88 [==============================] - 1s 6ms/step - loss: 0.1867 - accuracy: 0.9111 - val_loss: 0.6678 - val_accuracy: 0.8350
Epoch 76/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1720 - accuracy: 0.9254 - val_loss: 0.6959 - val_accuracy: 0.8100
Epoch 77/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1661 - accuracy: 0.9283 - val_loss: 0.7285 - val_accuracy: 0.8150
Epoch 78/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1665 - accuracy: 0.9240 - val_loss: 0.7048 - val_accuracy: 0.8150
Epoch 79/100
88/88 [==============================] - 1s 6ms/step - loss: 0.1734 - accuracy: 0.9226 - val_loss: 0.7252 - val_accuracy: 0.7950
Epoch 80/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1914 - accuracy: 0.9082 - val_loss: 0.7635 - val_accuracy: 0.7900
Epoch 81/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1683 - accuracy: 0.9269 - val_loss: 0.7489 - val_accuracy: 0.8250
Epoch 82/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1532 - accuracy: 0.9297 - val_loss: 0.9919 - val_accuracy: 0.8200
Epoch 83/100
88/88 [==============================] - 1s 6ms/step - loss: 0.2222 - accuracy: 0.8996 - val_loss: 0.7857 - val_accuracy: 0.8100
Epoch 84/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1493 - accuracy: 0.9362 - val_loss: 0.7353 - val_accuracy: 0.8300
Epoch 85/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1443 - accuracy: 0.9376 - val_loss: 0.8307 - val_accuracy: 0.8250
Epoch 86/100
88/88 [==============================] - 0s 6ms/step - loss: 0.1470 - accuracy: 0.9319 - val_loss: 0.9364 - val_accuracy: 0.8100
Epoch 87/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1273 - accuracy: 0.9427 - val_loss: 1.0125 - val_accuracy: 0.7900
Epoch 88/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1860 - accuracy: 0.9161 - val_loss: 0.6843 - val_accuracy: 0.7850
Epoch 89/100
88/88 [==============================] - 0s 3ms/step - loss: 0.1370 - accuracy: 0.9427 - val_loss: 0.9538 - val_accuracy: 0.8200
Epoch 90/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1337 - accuracy: 0.9362 - val_loss: 0.8614 - val_accuracy: 0.7900
Epoch 91/100
88/88 [==============================] - 1s 6ms/step - loss: 0.1467 - accuracy: 0.9405 - val_loss: 0.9644 - val_accuracy: 0.8000
Epoch 92/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1720 - accuracy: 0.9247 - val_loss: 0.7086 - val_accuracy: 0.7750
Epoch 93/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1470 - accuracy: 0.9419 - val_loss: 0.8167 - val_accuracy: 0.8000
Epoch 94/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1345 - accuracy: 0.9448 - val_loss: 0.8788 - val_accuracy: 0.8100
Epoch 95/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1211 - accuracy: 0.9505 - val_loss: 0.9519 - val_accuracy: 0.8200
Epoch 96/100
88/88 [==============================] - 0s 4ms/step - loss: 0.1266 - accuracy: 0.9498 - val_loss: 1.0115 - val_accuracy: 0.8450
Epoch 97/100
88/88 [==============================] - 1s 6ms/step - loss: 0.1249 - accuracy: 0.9513 - val_loss: 0.8444 - val_accuracy: 0.7900
Epoch 98/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1328 - accuracy: 0.9427 - val_loss: 0.8853 - val_accuracy: 0.8050
Epoch 99/100
88/88 [==============================] - 0s 5ms/step - loss: 0.1067 - accuracy: 0.9541 - val_loss: 0.8416 - val_accuracy: 0.7900
Epoch 100/100
88/88 [==============================] - 0s 6ms/step - loss: 0.1171 - accuracy: 0.9491 - val_loss: 0.9918 - val_accuracy: 0.8150

8. Evaluating Model

scores = model.evaluate(X_train,y_train, verbose=1)
print(scores)

44/44 [==============================] - 0s 2ms/step - loss: 0.0776 - accuracy: 0.9706
[0.07764449715614319, 0.9706093072891235]

9. Testing Model

y_pred=model.predict(X_test)
y_pred = (y_pred > 0.45)

score=accuracy_score(y_pred,y_test)
print(score)

0.7819548872180451

10. Confusion Matrix

conf_matrix = confusion_matrix(y_test, y_pred)
print(conf_matrix)

[[109  34]
 [ 53 203]]

fig, ax = plt.subplots(figsize=(7.5, 7.5))
ax.matshow(conf_matrix, cmap=plt.cm.Blues, alpha=0.3)
for i in range(conf_matrix.shape[0]):
    for j in range(conf_matrix.shape[1]):
        ax.text(x=j, y=i,s=conf_matrix[i, j], va='center', ha='center', size='xx-large')
 
plt.xlabel('Predictions', fontsize=18)
plt.ylabel('Actuals', fontsize=18)
plt.title('Confusion Matrix', fontsize=18)
plt.show()

print(classification_report(y_test,y_pred))

              precision    recall  f1-score   support

         0.0       0.67      0.76      0.71       143
         1.0       0.86      0.79      0.82       256

    accuracy                           0.78       399
   macro avg       0.76      0.78      0.77       399
weighted avg       0.79      0.78      0.78       399

11. summarize history for accuracy

plt.plot(model_his.history['accuracy'])
plt.plot(model_his.history['val_accuracy'])
plt.title('model accuracy')
plt.ylabel('accuracy')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()

12.summarize history for loss

plt.plot(model_his.history['loss'])
plt.plot(model_his.history['val_loss'])
plt.title('model loss')
plt.ylabel('loss')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()

13 Serialize model to JSON

model_json = model.to_json()
with open("Communities_Crime.json", "w") as json_file:
    json_file.write(model_json)
    
    

14. Serialize weights to HDF5

model.save_weights("Communities_Crime.h5")